I would suggest that you run 'netstat -s' on the netbsd box before and after the failed ls and diff them to see what counters are increasing. In particular see fragments received and reassembled. But saving them both to a file and diff is often very illuminating. Also, are you running a firewall that might perhaps be blocking fragments? Dump firewall counters before and after. Also, run tcpdump with -s1500 on both the freebsd and netbsd boxes, and see if there are any missing packets on one vs the other. It wasn't clear to me from your message that the dump was taken on the netbsd side. I have seen NFS fail (with the old 8K packets) on a network that was badly terminated (back in the 10base2 days) because the back-to-back packets self-collided while single packets with some spacing made it fine. This doesn't seem likely, but just barely worth mentioning.
Attachment:
pgpuQLUsTmFvc.pgp
Description: PGP signature